手写体识别

import paddle
from paddle.nn import Linear
import paddle.nn.functional as F
import os
import numpy as np
import matplotlib.pyplot as plt

train_dataset = paddle.vision.datasets.MNIST(mode='train')

train_data0 = np.array(train_dataset[0][0])
train_label_0 = np.array(train_dataset[0][1])


import matplotlib.pyplot as plt
plt.figure("Image") 
plt.figure(figsize=(2,2))
plt.imshow(train_data0, cmap=plt.cm.binary)
plt.axis('on') 
plt.title('image') 
plt.show()

print("图像数据形状和对应数据为:", train_data0.shape)
print("图像标签形状和对应数据为:", train_label_0.shape, train_label_0)
print("\n打印第一个batch的第一个图像，对应标签数字为{}".format(train_label_0))

<Figure size 432x288 with 0 Axes>

图像数据形状和对应数据为: (28, 28)
图像标签形状和对应数据为: (1,) [5]

打印第一个batch的第一个图像，对应标签数字为[5]

class MNIST(paddle.nn.Layer):
    def __init__(self):
        super(MNIST, self).__init__()
        
        self.fc = paddle.nn.Linear(in_features=784, out_features=1)

    def forward(self, inputs):
        outputs = self.fc(inputs)
        return outputs

model = MNIST()

def train(model):

    model.train()

    train_loader = paddle.io.DataLoader(paddle.vision.datasets.MNIST(mode='train'), 
                                        batch_size=16, 
                                        shuffle=True)

    opt = paddle.optimizer.SGD(learning_rate=0.001, parameters=model.parameters())

def norm_img(img):

    assert len(img.shape) == 3
    batch_size, img_h, img_w = img.shape[0], img.shape[1], img.shape[2]
 
    img = img / 255
    img = paddle.reshape(img, [batch_size, img_h*img_w])
    
    return img

import paddle
paddle.vision.set_image_backend('cv2')

model = MNIST()

def train(model):

    model.train()

    train_loader = paddle.io.DataLoader(paddle.vision.datasets.MNIST(mode='train'), 
                                        batch_size=16, 
                                        shuffle=True)

    opt = paddle.optimizer.SGD(learning_rate=0.001, parameters=model.parameters())
    EPOCH_NUM = 10
    for epoch in range(EPOCH_NUM):
        for batch_id, data in enumerate(train_loader()):
            images = norm_img(data[0]).astype('float32')
            labels = data[1].astype('float32')

            predicts = model(images)

            loss = F.square_error_cost(predicts, labels)
            avg_loss = paddle.mean(loss)

            if batch_id % 1000 == 0:
                print("epoch_id: {}, batch_id: {}, loss is: {}".format(epoch, batch_id, avg_loss.numpy()))
                
            avg_loss.backward()
            opt.step()
            opt.clear_grad()
            
train(model)
paddle.save(model.state_dict(), './mnist.pdparams')

epoch_id: 0, batch_id: 0, loss is: [27.441448]
epoch_id: 0, batch_id: 1000, loss is: [2.281343]
epoch_id: 0, batch_id: 2000, loss is: [4.0873246]
epoch_id: 0, batch_id: 3000, loss is: [4.101654]
epoch_id: 1, batch_id: 0, loss is: [4.409295]
epoch_id: 1, batch_id: 1000, loss is: [2.039107]
epoch_id: 1, batch_id: 2000, loss is: [2.5832603]
epoch_id: 1, batch_id: 3000, loss is: [2.6795988]
epoch_id: 2, batch_id: 0, loss is: [3.9104843]
epoch_id: 2, batch_id: 1000, loss is: [3.9539223]
epoch_id: 2, batch_id: 2000, loss is: [3.0901222]
epoch_id: 2, batch_id: 3000, loss is: [2.88086]
epoch_id: 3, batch_id: 0, loss is: [2.8132215]
epoch_id: 3, batch_id: 1000, loss is: [3.9469924]
epoch_id: 3, batch_id: 2000, loss is: [2.484096]
epoch_id: 3, batch_id: 3000, loss is: [2.9386535]
epoch_id: 4, batch_id: 0, loss is: [4.952403]
epoch_id: 4, batch_id: 1000, loss is: [4.189546]
epoch_id: 4, batch_id: 2000, loss is: [4.81828]
epoch_id: 4, batch_id: 3000, loss is: [2.296103]
epoch_id: 5, batch_id: 0, loss is: [5.300122]
epoch_id: 5, batch_id: 1000, loss is: [3.3335328]
epoch_id: 5, batch_id: 2000, loss is: [1.3920281]
epoch_id: 5, batch_id: 3000, loss is: [4.3053403]
epoch_id: 6, batch_id: 0, loss is: [3.0284095]
epoch_id: 6, batch_id: 1000, loss is: [1.5199479]
epoch_id: 6, batch_id: 2000, loss is: [3.2134027]
epoch_id: 6, batch_id: 3000, loss is: [2.2408743]
epoch_id: 7, batch_id: 0, loss is: [2.8081536]
epoch_id: 7, batch_id: 1000, loss is: [1.917313]
epoch_id: 7, batch_id: 2000, loss is: [5.512356]
epoch_id: 7, batch_id: 3000, loss is: [3.0259757]
epoch_id: 8, batch_id: 0, loss is: [2.2609704]
epoch_id: 8, batch_id: 1000, loss is: [2.1595194]
epoch_id: 8, batch_id: 2000, loss is: [3.7097778]
epoch_id: 8, batch_id: 3000, loss is: [4.8904915]
epoch_id: 9, batch_id: 0, loss is: [2.3321204]
epoch_id: 9, batch_id: 1000, loss is: [5.85237]
epoch_id: 9, batch_id: 2000, loss is: [4.1020594]
epoch_id: 9, batch_id: 3000, loss is: [1.5817034]

import matplotlib.pyplot as plt
import numpy as np
from PIL import Image

img_path = './work/example_0.jpg'

im = Image.open('./work/example_0.jpg')
plt.imshow(im)
plt.show()

im = im.convert('L')
print('原始图像shape: ', np.array(im).shape)

im = im.resize((28, 28), Image.ANTIALIAS)
plt.imshow(im)
plt.show()
print("采样后图片shape: ", np.array(im).shape)

原始图像shape:  (252, 255)

采样后图片shape:  (28, 28)

def load_image(img_path):
    im = Image.open(img_path).convert('L')

    im = im.resize((28, 28), Image.ANTIALIAS)
    im = np.array(im).reshape(1, -1).astype(np.float32)

    im = 1 - im / 255
    return im

model = MNIST()
params_file_path = 'mnist.pdparams'
img_path = './work/example_0.jpg'

param_dict = paddle.load(params_file_path)
model.load_dict(param_dict)

model.eval()
tensor_img = load_image(img_path)
result = model(paddle.to_tensor(tensor_img))
print('result',result)
print("本次预测的数字是", result.numpy().astype('int32'))

result Tensor(shape=[1, 1], dtype=float32, place=CPUPlace, stop_gradient=False,
       [[3.76953363]])
本次预测的数字是 [[3]]